Calculate Ammonia (NH3) Emissions

[jrissman]

According to Jonathan Buonocore:

[Ammonia is] a major precursor to PM, especially from agriculture, and is a major limiting reagent in cities. VOCs is also a fairly hefty precursor to secondary organic aerosols, which is relevant for some sectors,

According to this paper by the USDA, 60-85% of ammonia emissions come from agricultural sources. Therefore, policies that affect agriculture (such as the lever that shifts some demand from animal to non-animal products) may have a large impact on ammonia emissions.

No public health incidence-per-ton information is available directly from EPA to estimate the health outcomes of ammonia emissions. Having ammonia emissions data would mostly be useful only (1) if feeding EPS results into a gridded air quality model like GeosCHEM or CMAQ, or (2) if we commission someone to produce health impact-per-ton multipliers, including a multiplier for ammonia.

Suggested by Jonathan Buonocore on 7/10/2020

[jrissman]

For the U.S., the 2017 National Emissions Inventory (released April 2020) contains Ammonia emissions broken out into source types. Go here and click on the "Data Queries" tab, then in part 1, select "Ammonia" (leaving other settings alone, such as "National" scope) and click "Submit". The resulting data are in the Excel file attached to this comment here:

2020-07-15 US NEI 2017 Ammonia Emissions.xlsx

Here is a summary breakdown of the ammonia emissions categorized by EPS component:

EPS Category	Tons	Percent
Agriculture	3558159	82.4%
LULUCF	485731	11.2%
Transportation	103511	2.4%
Buildings	54192	1.3%
Industry (process emissions)	80605	1.9%
Electricity	19800	0.5%
Industry (fuel use)	12542	0.3%
Other	5409	0.1%

LULUCF is almost all fires, mostly wildfires, and can be excluded as non-anthropogenic if desired. (There are some prescribed burns, but there would presumably be more wildfire emissions without the prescribed burns, so we should probably treat these two categories the same: include or exclude.) If we exclude all LULUCF, agriculture rises to a 93% share of the total, with the rest spread thinly across all the other sectors. So it would probably be fair to only include agricultural ammonia emissions, rather than add a lot of structure and detail to try to capture that last 7% spread across the other sectors.

I don't think adding agricultural ammonia emissions would be too much work.

[jrissman]

The 2017 National Emissions Inventory technical support document cites (on page 4-191) the following document for its ammonia emissions indices for residential fuel combustion:

U.S. Environmental Protection Agency. 2004. Emission Inventory Improvement Program. Estimating Ammonia Emissions from Anthropogenic Sources, Draft Final Report. Prepared by E.H. Pechan and Associates, Inc. Research Triangle Park, NC. https://www.epa.gov/sites/production/files/2015-08/documents/eiip_areasourcesnh3.pdf

That document also contains ammonia emissions indices for other combustion sources, such as utilities, commercial and institutional buildings, etc. by fuel type (on pages 32-33) and for mobile sources (on-road page 41, non-road pages 44-45). This document might be a good source for emissions indices if we decide to include ammonia emissions from combustion (as a regular, 13th pollutant type) and not just from agriculture.

Jonathan Buonocore suggests that ammonia emissions from buildings can have disproportionate impact on health outcomes because ammonia is often a limiting reagent for PM formation in cities, so there might be a reason to model the ammonia emissions from non-agriculture sources even though those ammonia emissions are so much smaller than ammonia emissions from agriculture.